BMC Public Health

12 Pages·2008·0.36 MB·English

BMC Public Health
BioMed Central
Research article Open Access
Physical activity and dietary behaviour in a population-based sample
of British 10-year old children: the SPEEDY study (Sport, Physical
activity and Eating behaviour: Environmental Determinants in
Young people)
Esther MF van Sluijs*1, Paula ML Skidmore2, Kim Mwanza2,
Andrew P Jones3, Alison M Callaghan1, Ulf Ekelund1, Flo Harrison3,
Ian Harvey2, Jenna Panter3, Nicolas J Wareham1, Aedin Cassidy2 and
Simon J Griffin1
Address: 1Medical Research Council Epidemiology Unit, Institute of Metabolic Science, Box 285, Addenbrookes Hospital, Hills Road, Cambridge,
CB2 0QQ, UK, 2School of Medicine, Health Policy and Practice, University of East Anglia, Norwich, NR4 7JT, UK and 3School of Environmental
Sciences, University of East Anglia, Norwich, NR4 7JT, UK
Email: EstherMFvan Sluijs*[email protected]; [email protected];
[email protected]; [email protected]; [email protected];
[email protected]; [email protected]; [email protected];
[email protected]; [email protected]; [email protected];
[email protected]
* Corresponding author
Published: 14 November 2008 Received: 9 June 2008
Accepted: 14 November 2008
BMC Public Health 2008, 8:388 doi:10.1186/1471-2458-8-388
This article is available from: http://www.biomedcentral.com/1471-2458/8/388
© 2008 van Sluijs et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
Background: The SPEEDY study was set up to quantify levels of physical activity (PA) and dietary
habits and the association with potential correlates in 9–10 year old British school children. We
present here the analyses of the PA, dietary and anthropometry data.
Methods: In a cross-sectional study of 2064 children (926 boys, 1138 girls) in Norfolk, England,
we collected anthropometry data at school using standardised procedures. Body mass index (BMI)
was used to define obesity status. PA was assessed with the Actigraph accelerometer over 7 days.
A cut-off of ≥ 2000 activity counts was used to define minutes of moderate-to-vigorous PA
(MVPA). Dietary habits were assessed using the Health Behaviour in School Children food
questionnaire. Weight status was defined using published international cut-offs (Cole, 2000).
Differences between groups were assessed using independent t-tests for continuous data and chi-
squared tests for categorical data.
Results: Valid PA data (>500 minutes per day on ≥ 3 days) was available for 1888 children. Mean
(± SD) activity counts per minute among boys and girls were 716.5 ± 220.2 and 635.6 ± 210.6,
respectively (p < 0.001). Boys spent an average of 84.1 ± 25.9 minutes in MVPA per day compared
to 66.1 ± 20.8 among girls (p < 0.001), with an average of 69.1% of children accumulating 60 minutes
each day. The proportion of children classified as overweight and obese was 15.0% and 4.1% for
boys and 19.3% and 6.6% for girls, respectively (p = 0.001). Daily consumption of at least one
Page 1 of 12
(page number not for citation purposes)
BMC Public Health 2008, 8:388 http://www.biomedcentral.com/1471-2458/8/388
portion of fruit and of vegetables was 56.8% and 49.9% respectively, with higher daily consumption
in girls than boys and in children from higher socioeconomic backgrounds.
Conclusion: Results indicate that almost 70% of children meet national PA guidelines, indicating
that a prevention of decline, rather than increasing physical activity levels, might be an appropriate
intervention target. Promotion of daily fruit and vegetable intake in this age group is also warranted,
possibly focussing on children from lower socioeconomic backgrounds.
Background The SPEEDY study (Sport, Physical activity and Eating
A lack of physical activity and poor dietary habits are behaviour: Environmental Determinants in Young peo-
believed to be the major contributors to the current rise in ple) was established to examine physical activity levels
childhood obesity [1-3]. Both behaviours have also been and dietary behaviour in a large population-based sample
shown to have independent associations with other of British 9–10 year old children, and to investigate the
health problems during childhood, including metabolic individual and collective factors associated with these
impairments [4-8] and poor skeletal health [9,10]. They behaviours. The first aim of the SPEEDY study is to assess
also tend to co-exist both in adults and children [11,12]. the extent of the problem of physical inactivity and
Increasing physical activity and improving dietary habits unhealthy dietary habits. With its population-based sam-
in childhood have therefore been identified as targets for pling strategy, large sample size and detailed assessment
future public health policy [1,13]. To be able to effectively of both physical activity and diet, the SPEEDY study will
promote changes in these complex health behaviours, add to the existing evidence base. The aim of the current
reliable and valid data are needed about current patterns, paper is to describe the SPEEDY study and the levels of
and the demographic and modifiable factors that are most physical activity and dietary habits of its participants.
strongly associated with them. This will aid the identifica-
tion of populations at risk and the development of inter- Methods
ventions to promote change in these health behaviours. Study sample
School recruitment
To date, the extent of the problem of physical inactivity Schools in the county of Norfolk, South-East England,
and unhealthy dietary habits in children is largely were sampled purposively to achieve heterogeneity in
unknown. Descriptive data on patterns and levels of phys- location. For logistical purposes, only schools with at least
ical activity assessed using valid and reliable measures in 12 Year 5 pupils (aged 9–10 years) were sampled. Head
large population-based samples is scarce, especially in teachers at selected schools were first sent an invitation
children [14]. Previous studies have reported mixed letter and an information leaflet, and then contacted to
results, with the proportion of children achieving the rec- arrange a face-to-face visit. In this 10-minute visit, details
ommended guideline of at least one hour of moderate-to- were provided about the aims of the study, the invitation
vigorous intensity physical activity (MVPA) each day [13] and measurement procedures, the degree of school
varying from 2.5 to 97% [13,15-18]. This large variation involvement required, and potential benefits of participa-
mainly depends on the population studied and the varia- tion. An information pack was provided at the visit, con-
tion in data processing methods applied. However, most taining a copy of the ethical approval, examples of all
studies indicate that physical activity levels decline with study materials, and information leaflets for all Year 5
age, especially during late primary school years and teachers. If they agreed that their school could take part,
throughout secondary school [19-21], making this a head teachers were asked to send the study team a signed
potentially important period for health promotion inter- acceptance letter. Upon request, a smaller sister school of
ventions. Although there is some information available one of the participating schools (a smaller school run
on the nutritional content of the diet of primary school- under shared administration) was allowed to participate.
aged children (e.g. [22,23]), little is known about their School recruitment was completed prior to the start of
dietary behaviour such as food choice. Previous studies data collection in April 2007.
either primarily focussed on secondary school-aged chil-
dren [24] or were conducted a decade ago [25]. In order to Participant recruitment
develop behavioural interventions targeted at improving Ethical approval for the SPEEDY study was obtained from
dietary habits and potentially prevent obesity in primary the University of East Anglia local research ethics commit-
school-aged children, up-to-date information about the tee. Teams of two research assistants visited participating
dietary habits is required. schools two weeks before the measurement date and
introduced the SPEEDY study to all Year 5 children. Chil-
dren were given an information pack containing a leaflet
Page 2 of 12
(page number not for citation purposes)
BMC Public Health 2008, 8:388 http://www.biomedcentral.com/1471-2458/8/388
for themselves, a letter for their parents/guardians, and a directly obtained from schools. Urban/rural classification
consent form. Children were instructed to return the com- for the schools was determined using Bibby and Shep-
pleted consent form to the school if they were willing to herd's (2004) classification of rurality [27] and four den-
participate. A classroom poster was left at the school to sity profiles, ranging from 'hamlet & isolated dwelling' to
remind the children of when the data collection was tak- 'urban, more than 10,000 inhabitants', were used in the
ing place. Three days before the measurement visit, a present study.
member of the research team contacted the school to
assess response rates and to ask the class teacher to rein- Demographic information
force the message that children would not be able to par- Children self-reported their date of birth, and age was cal-
ticipate without completed consent. Only children with a culated using the measurement date. Ethnicity was
fully completed consent form (signed by both a parent/ reported separately by both parents using the UK standard
guardian and the child) on the day of measurement were classification. Results were combined to determine the
included in the study. An average of 168 children (range: child's ethnicity, which was collapsed into a dichotomous
121 to 242 per week) was recruited into the study each variable (white vs. other ethnic background). The parent
week with an aim of recruiting 2000 children over a 12- or main carer self-reported their highest educational qual-
week period. ifications (in categories) and home postcode to determine
urban/rural location.
Measurement procedures
Data collection was performed during the Summer term Anthropometry
of 2007 (April to July). Teams of two or more trained Simple non-invasive anthropometry measures were con-
research assistants conducted measurements at participat- ducted using standardized procedures with children
ing schools according to standard operating procedures. dressed in light clothing. Portable Leicester height meas-
They visited between five and 10 schools each week. All ures were used to measure height to the nearest millime-
participating children self-completed their questionnaires tre. Waist circumference was measured twice to the nearest
with at least one research assistant available for assistance. millimetre at the midpoint between the lower costal mar-
Completeness of the questionnaire was determined on- gin and the level of the anterior superior iliac crests, using
site for quality control. Concurrently, research assistants a Seca 200 measuring tape. A third measurement was
measured anthropometry, obtained a saliva sample (chil- taken if a discrepancy of three or more centimetres was
dren were asked to chew on a cotton swab) and explained observed and an average was calculated. A 0.5 cm correc-
the contents of the home pack (containing the 4-day food tion was applied to account for clothing [28]. A non-seg-
diary, Actigraph accelerometer and parent questionnaire). mental bio-impedance scale (Tanita, type TBF-300A) was
Children were instructed to return the home pack to used to measure weight (to the nearest 0.1 kilogram) and
school eight days after the measurement day for collection impedance. Previously validated and published equations
by a member of the study team. Class teachers distributed were used to calculate body fat percentage (BF%), fat mass
a reminder letter for parents two days before pack collec- (FM) and fat free mass (FFM) from the impedance value
tion. Completion of packs was checked upon collection [29]. Body mass index (BMI) was calculated as weight (in
and follow-up of unreturned packs was conducted by the kilograms) divided by height squared (in meters). Obesity
school administrative staff. status was determined using gender- and age-dependent
cut points [30]. All scales were calibrated before and half-
Data collection and processing way through data collection. Quality assurance was estab-
The descriptions below relate specifically to the measures lished by assessing inter-observer variability of height and
for which data are presented in this paper. waist circumference before and after data collection,
which was found to be acceptable (ranging from 0.1–0.3
Aggregated school-level information centimetres for height and 0.7–1.3 centimetres for waist
Norfolk County Council provided school-based summary circumference).
data on all Local Authority (state) schools. This included
school status (state vs. private), postcode, number of Year Physical activity
5 pupils, proportion of all children receiving free school Free-living physical activity was assessed over one week
meals, ethnicity distribution and participation in the with the ActiGraph activity monitor (GT1M, Actigraph
Healthy Norfolk Schools Programme [26]. Also provided LCC, Pensacola, US). The children wore the accelerometer
was data from the 2007 Norfolk Primary Care Trust Child on an elastic waistband on the right hip during waking
Height and Weight Survey, containing information on hours, except whilst bathing and during other aquatic
anthropometry for over 85% of Norfolk Year 6 children activities. Children kept an Actigraph diary to report when
(aged 10–11 years). Data on independent private schools they had taken the monitor off and for what reason. Activ-
was provided by the Independent Schools Council or ity data were stored at 5-second intervals and were down-
Page 3 of 12
(page number not for citation purposes)
BMC Public Health 2008, 8:388 http://www.biomedcentral.com/1471-2458/8/388
loaded to a computer upon receipt. In case of monitor Results
failure or lack of data, children were asked to re-wear the Recruitment
monitor for a further week. All monitors were calibrated Figure 1 provides an overview of recruitment of schools
using a mechanical spinner before the start of data collec- and children. A total of 227 (214 state and 13 independ-
tion. A bespoke programme (MAHUffe, http://www.mrc- ent) schools were eligible for recruitment, of which 157
epid.cam.ac.uk) was used for data reduction and further were sent an invitation letter. 101 Schools agreed to par-
analyses. The first day of data collection (day of measure- ticipate in the SPEEDY study and measurements sessions
ment at school) was removed from all files and 10 min- were conducted at 92 schools (58.6% of invited schools).
utes of continuous zero counts were classified as 'non At these 92 schools, 3619 Year 5 children were invited to
worn time'. The outcome variables were daily activity take part (range per school: 5–155), measurements were
counts per minute (cpm) and time (min) spent in moder- performed on a total of 2064 children (57.0% of eligible
ate to vigorous intensity physical activity (MVPA, >2000 sample) with response rates per school ranging from 13%
cpm). This threshold corresponds to a walking pace of to 100%. Of these, a total of 2043 children returned the
about 4 km/h in children [31], and has been applied suc- Actigraph accelerometer (99.0%), with 1868 (90.5%)
cessfully in this age group before to study associations providing valid data. No differences were observed
between physical activity intensity and metabolic out- between children with and without valid physical activity
comes [32]. Daily cpm is an indicator of the total volume data for BMI, overweight status, or parental education.
of physical activity (i.e. average intensity of physical activ- However, girls were more likely to not provide valid data.
ity). This variable was derived by dividing total counts by
monitoring time per day (between 6.00 and 23.00) and Sample description
averaged over the measurement period. We have previ- Table 1 provides an overview of the characteristics of par-
ously shown that this variable is significantly correlated ticipating schools. Compared with all eligible Norfolk
with physical activity energy expenditure obtained by the schools, village schools were over-sampled in the SPEEDY
doubly-labelled water method [33]. Children who did not study, with an under-sampling of schools in urban areas
manage to record valid data for at least 500 minutes per and of independent schools. Table 2 provides a descrip-
day on at least 3 days were excluded from further analyses. tion of characteristics of the SPEEDY children. A higher
proportion of girls than boys participated and only a
Dietary behaviour small proportion of children were from a non-white eth-
Food choice was assessed using an adapted version of the nic background.
Health Behaviour in School Children (HBSC) question-
naire, a 15-item questionnaire of the most commonly Anthropometry
consumed foods that has been previously validated in a Table 2 describes the anthropometric data collected in the
Belgian sample and can be used for ranking participants SPEEDY study. Compared with boys, girls were heavier,
for most food items [34]. Adaptations to the items had less fat free mass and more fat mass, a higher body fat
included removing alcohol consumption, adding fruit percentage and a higher BMI (all p < 0.001). In addition,
juice consumption and separating sweets and chocolate a higher proportion of girls than boys were classified as
into two separate items. Children reported on the fre- either overweight or obese.
quency of consumption on a 7-point scale. Results were
dichotomized at less than once per day ("never", "less Physical activity
than once per week", "once per week", "2–4 days per Table 3 describes the data on objectively measured physi-
week", "5–6 days per week") vs. once or more per day cal activity stratified by gender. Overall physical activity
("once a day, every day", "every day, more than once") and time (min/day) spent in MVPA was significantly
[35]. higher (p < 0.001) in boys than girls. Both boys and girls
did more activity overall on weekend days compared with
Statistical analyses weekdays. However, time spent at moderate and vigorous
Descriptive data were summarized as means with stand- intensity activity did not differ between weekend and
ard deviations or percentages, and differences between weekdays in either gender. Figure 2 shows that a higher
groups were tested using independent samples t-tests or proportion of boys than girls met the physical activity
Pearson Chi-square tests. Differences between activity lev- guideline, and that the proportion differs by weight status
els on week- and weekend days were tested using paired t- and home location
tests, and Chi-square tests were used to test for differences
between groups meeting the physical activity guideline or Dietary behaviour
consuming at least one portion of fruit and vegetable per Table 4 shows the results of the food choice question-
day. naire. Daily consumption of fruit and of vegetables was
56.8% and 49.9% respectively, with girls reporting daily
Page 4 of 12
(page number not for citation purposes)
BMC Public Health 2008, 8:388 http://www.biomedcentral.com/1471-2458/8/388
282 schools in Norfolk with Year 5 pupils on
role
55 schools excluded
Reason: less than 12 Year 5 pupils on role
227 eligible schools
70 schools not sampled
157 schools sampled to be approached (via
post and telephone follow up)
65 schools did not take part
Reasons: refused visit (N=20); unavailable for visit
(N=33); refused after visit (N=4); waiting list (N=8)
92 schools visited for measurements (58.6%)

3619 available Year 5 pupils invited
1552 children did not participate
Reasons: not available on day (N=64); declined
(N=282); consent not returned (N=1206)
2067 year 5 children consented and
participated in measurements (57.1%)
3 children withdrew after measurements
2064 participants available for analyses (57.0%
of eligible sample)
FFliogwucrhea 1rt of recruitment of schools and children into the SPEEDY study
Flowchart of recruitment of schools and children into the SPEEDY study.
Page 5 of 12
(page number not for citation purposes)
BMC Public Health 2008, 8:388 http://www.biomedcentral.com/1471-2458/8/388
Table 1: Characteristics of the schools participating in SPEEDY study compared with available data on eligible Norfolk schools.
SPEEDY schools All eligible Norfolk schools
School location
- Urban (>10,000 inhabitants) 32.6% 34.7%
- Town and fringe 39.1% 38.0%
- Village 23.9% 20.7%
- Hamlet & isolated dwelling 4.3% 6.6%
Independent school 3.3% 7.5%
Free school meals (% children) (median) 9.0% 9.0%
Norfolk Healthy Schools participation (%) 38.2% 42.6%
Ethnicity (% non-British White) (median) 5.2% 4.6%
Weight status (%)*
- overweight 13.4% 13.8%
- obese 16.7% 15.8%
No of year 5 pupils (median) 31 31
Sex (% boys in Year 6)* 52.5% 51.0%
NOTE: Figures based on data aggregated at school level. Data on Healthy Norfolk Participation, ethnicity, BMI and sex were only available for state
schools, %Free school meals not applicable for independent schools.
*: assessed in Summer 2007 in Year 6 children (Norfolk Primary Care Trust child Height and Weight Survey).
consumption more frequently than boys. 41.3% of chil- obese children, whereas children from a higher socioeco-
dren reported eating at least one piece of fruit and one nomic background were more likely to eat fruit and vege-
serving of vegetables per day, with significant differences table daily.
between sexes (see Figure 2). This percentage also
increased with increasing socioeconomic status, but did Physical activity
not differ by weight status or home location. It was also Estimates of total physical activity, represented as average
not associated with meeting physical activity guidelines counts per minute, are largely comparable to the results of
(40.7% vs. 42.2% for whether or not participants were previous population-based studies in a similar age group
meeting the physical activity guideline, respectively). in Britain and other European countries [15,16,36,37].
More than half of children reported daily consumption of The small differences observed may be due to differences
skimmed or semi-skimmed milk, breakfast cereals and in monitors used and data processing strategies. However,
fruit juices. it may also be due to the known age-related decline in
physical activity [16,19], potential decreases in physical
Discussion activity levels in the past decade [15,37], or seasonal
The aim of the current paper was to present data on phys- effects [15,16,36]. As consistently reported before
ical activity and dietary behaviour from a population- [15,16,19], boys were substantially more active than girls.
based sample of British 9–10 year old children. Results In contrast to previous work, the data in this study were
showed that more than two-thirds of children adhere to collected only during the spring and early summer. A sea-
the physical activity guideline of accumulating at least 60 sonal effect on physical activity levels in children has pre-
minutes of MVPA each day, but that daily consumption of viously been established [16,38,39], with lower activity
fruit or of vegetables was only reported by 56.8% and levels observed during cold or extreme weather. Therefore,
49.9% of the children. Boys were more likely to be physi- our estimate of overall physical activity may be overesti-
cally active and of normal weight than girls. In contrast, mated compared with studies collecting data throughout
boys' reported daily consumption of 'healthy' foods was the year. The timing of our measurements may also
lower, and their consumption of snacks and unhealthy explain the observation that children were consistently
food items such as soft drinks was higher than that of girls. more active on weekend days than on weekdays, in con-
In addition, normal weight children were more likely to trast to previous work [16,40]. During European spring
meet physical activity guidelines than overweight and and summer, children will be more likely to play outside
Page 6 of 12
(page number not for citation purposes)
BMC Public Health 2008, 8:388 http://www.biomedcentral.com/1471-2458/8/388
for longer periods of time in their free time as a result of
Table 2: Descriptive personal and anthropometry data on the
comfortable temperatures and longer hours of daylight.
SPEEDY study sample (all numbers in cells are mean (standard
deviation) unless stated otherwise) Time spent outside has previously been associated with
higher activity levels [19].
Boys Girls Total
In contrast to total activity levels, estimates of time spent
N (%) 926 (44.9) 1138 (55.1) 2064
in at least moderate-to-vigorous physical activity and of
proportions of children meeting guidelines differ from
Age (years) 10.2 (0.3) 10.3 (0.3) 10.3 (0.3)
other studies. Previous estimates of minutes of MVPA per
day in this age group range from 20 to 192 minutes per
Ethnicity (% non-white) 3.8 3.8 3.8
day, with 2.5 to 97% achieving the recommended guide-
line of at least 60 minutes of MVPA each day of the week
SES – parental education (%)
- GCSE or lower 36.3 40.6 38.7 [13,15-18]. On average, the children in the SPEEDY study
- up to A level 42.2 40.4 41.2 accumulated 74.1 minutes of MVPA per day, with 69.1%
- higher education 21.4 18.9 20.0 of children being classified as sufficiently active according
to current physical activity recommendations. This value
Home location is higher compared with previous estimates in British chil-
- Urban (>10,000 inhabitants) 40.3 38.9 39.6
dren recently reported [16]. However, data on time spent
- Town and fringe 26.6 29.9 28.4
at MVPA and prevalence of children being sufficiently
- Village 25.4 24.7 25.0
active according to recommendations are not directly
- Hamlet & isolated dwelling 7.7 6.5 7.0
comparable between studies. This is explained by differ-
Height (cm) 140.9 (6.5) 141.0 (6.8) 141.0 (6.7) ent thresholds used when defining MVPA, for which cur-
rently no consensus exists. Previously published
Weight (kg) 35.7 (7.7) 37.1 (8.8)* 36.5 (8.4) thresholds for MVPA vary substantially depending on the
activities performed during the calibration process, and
Bodyfat% 27.2 (7.5) 33.2 (7.3)* 30.5 (7.9) the choice of intensity threshold will affect the results for
MVPA substantially. This at least in part explains the dis-
Waist circumference (cm) 63.6 (7.8) 63.6 (9.0) 63.6 (8.4) crepancy with previously published British data [16],
where the authors used a threshold of 3600 cpm, a sub-
BMI (kg/m2) 17.9 (2.9) 18.5 (3.4)* 18.2 (3.2) stantially higher threshold than used in previous work
[15,17] and in the current study. This threshold (2000
Weight status (%)
cpm), however, lies mid-way between the various inten-
- overweight 15.0 19.3* 17.4
sity thresholds derived in children, which range from 615
- obese 4.1 6.6 5.5
[41] to 3600 counts per minute [16].
*Statistical significant difference between boys and girls at p ≤ 0.001
One interpretation of our data could be that only one
third of children at this age are insufficiently active. How-
ever, this interpretation may be dangerous as recent data
Table 3: Description of physical activity variables as collected with Actigraph accelerometers in the SPEEDY study.
Boys (N = 822) Girls (N = 1046) Total (N = 1868)
Counts per minute
- Weekday 682.6 (176.9) 592.9 (163.5)* 632.4 (175.3)
- Weekend day 754.8 (346.9)# 690.7 (343.9)*,# 719.6 (346.7)#
- Total week 716.5 (220.2) 635.6 (210.6)* 671.2 (218.6)
Minutes of MVPA
- Weekday 83.5 (23.4) 65.5 (19.5)* 73.5 (23.1)
- Weekend day 85.1 (38.5) 67.2 (30.2)* 75.3 (35.3)
- Total week 84.1 (25.9) 66.1 (20.8)* 74.1 (24.9)
Differences between boys and girls tested with independent t-tests, differences between week and weekend day activity with paired samples t-tests.
*: difference between boys and girls at P < 0.001.
#: difference between weekday and weekend day activity within group at p < 0.001
Page 7 of 12
(page number not for citation purposes)
BMC Public Health 2008, 8:388 http://www.biomedcentral.com/1471-2458/8/388
A.
Boys Girls
p<0.001
90
meeting target678000 p<0.001
Percentage of children 1234500000
0
60 mins PA Daily fruit and vegetable

B.
Normal weight Overweight Obese
90
p<0.001
meeting target678000
Percentage of children 1234500000
0
60 mins PA Daily fruit and vegetable

C.
GCSE or lower Up to A-level Higher education
90
meeting target678000 p<0.001
Percentage of children 1234500000
0
60 mins PA Daily fruit and vegetable

D.
Urban >10,000 inhabitants Town and fringe Village Hamlet
90
meeting target678000 p<0.05
Percentage of children 1234500000
0
60 mins PA Daily fruit and vegetable

aPtFinecigdsr cu (reArenep.t osa2egrexti ;n oBgf .c SwoPneEisEguDhmtY isn tcgah taiultds rl;e eCans. tmp oaerneeetn inptagiel pceehd yuoscfi cafatriluo aintc at(inivndidt yoic n(aPetAo sr)e ogrvuf iinsdoge cloiinof eev ceoogf en6to0am bmlieci npsuteatret usd sap)y;e ,a rs ntddraa yDt iof.i enhd oa mbny ea dvleoermcaagoteigo rdnaa)pyh iocf cthhaer awceteerki s-
Percentage of SPEEDY children meeting physical activity (PA) guideline of 60 minutes per day on an average
day of the week and reporting consuming at least one piece of fruit and one serving of vegetable per day, strat-
ified by demographic characteristics (A. sex; B. weight status; C. parental education (indicator of socioeco-
nomic status); and D. home location).
Page 8 of 12
(page number not for citation purposes)
BMC Public Health 2008, 8:388 http://www.biomedcentral.com/1471-2458/8/388
Table 4: Dietary habits of the SPEEDY sample, as measured with countries. Here daily fruit and vegetable consumption
food frequency questionnaire (numbers in cells are percentage ranged between 20–50% and 10–50%, respectively [24].
of children reporting to consume the food type at least once a
Results from 6081 participating English adolescents indi-
day)
cated that only 25% reported daily fruit consumption,
Boys Girls Total with daily vegetable consumption just under 30%. In
addition, about 40% of adolescents reported daily con-
Fruit 51.6 61.1* 56.8 sumption of soft drinks, compared with 12% in the
SPEEDY population. Whether these differences are due to
Vegetables 44.8 54.1* 49.9 age-related decline in healthy dietary habits, changes over
time (the HSBC data was collected in 2002), reporting
Sweets 18.4 12.1* 14.9
bias, or differences in populations studied, remains uncer-
tain.
Chocolate 18.6 14.2* 16.2
Results of the current study are largely in agreement with
Sugary soft drinks (e.g. coke) 16.1 9.3* 12.4
results from the National Diet and Nutrition Survey
(NDNS) in children aged 7 to 10 years, conducted in 1997
Diet soft drinks 16.2 11.7* 13.8
[25]. In this study, for example, consumption of semi- and
Skimmed/semi-skimmed milk 54.1 51.0 52.4 skimmed milk and white bread was also higher than
whole milk and brown or wholemeal bread. Moreover,
Whole fat milk 15.6 13.2 14.3 boys were less likely to regularly consume vegetables,
whereas girls consumed less breakfast cereals than boys.
Cheese 17.4 16.4 16.9 In contrast, the children in the SPEEDY study reported a
higher consumption of both artificially sweetened and
Breakfast cereals 60.1 51.9* 55.6 sugary soft drinks, which may reflect changes in children's
food consumption over time.
White bread 44.0 41.7 42.7
It is known that children's recall of food intake is prone to
Brown bread 22.3 20.0 21.0
reporting errors [42,43] and that children's interest and
enthusiasm may wane during lengthy dietary assessment
Crisps 24.1 20.5 22.1
periods. Therefore the short HBSC questionnaire was used
to assess intake of the most commonly consumed foods
Chips 7.8 5.3* 6.4
rather than exact intake of nutrients, in order to collate
information on trends in dietary intake. Under-reporting
Fruit juices 54.5 54.9 54.7
of energy intake has previously been reported in both
boys and girls [44,45], with under-reporters reporting
Differences between boys and girls tested with Pearson Chi-square.
*: differences between boys and girls at p < 0.05 consuming the same number of fruit and vegetable serv-
ings, but fewer servings from sweets and fats groups [44].
suggest that about 90 to 120 minutes of moderate inten- Under-reporting is also associated with weight status [46],
sity physical activity is needed to prevent clustering of car- possibly explaining the observed healthier dietary profile
diovascular risk factors in children [6]. Given the well- in girls compared to boys given the differences in weight
established age-related decline in physical activity which status by gender. As with the physical activity measure-
appears to start before puberty [20,21], examining the cor- ment, it is likely that seasonal influences have played a
relates and determinants of physical activity and of role in the assessment of dietary intake [47]. In particular
changes in physical activity is therefore of particular inter- fruit and vegetable availability and choice of food and
est in this age group in order to inform preventive action. drink consumption in warmer weather might have
Such interventions may be better placed if they focus on increased the reported consumption frequency of fruit,
preventing declines in physical activity, and not necessar- vegetables and soft drinks.
ily increases in physical activity.
Recruitment and representativeness
Diet One of the aims of the sampling strategy in the SPEEDY
A pattern of unhealthy food choices was revealed with study was to achieve maximum environmental heteroge-
high consumption of energy and fat dense food and neity in order to address questions relating to environ-
refined carbohydrates. Our results however compare mental influences on behaviour. In line with this we
favourably to those from a previously reported HBSC sur- recruited a larger proportion of schools in rural than in
vey in a slightly older population (11–15 year olds) in 35 urban locations. This might have affected the prevalence
Page 9 of 12
(page number not for citation purposes)
BMC Public Health 2008, 8:388 http://www.biomedcentral.com/1471-2458/8/388
of certain behaviours amongst our school sample, and lates from the psychological, biological, socio-cultural
hence their representativeness to the wider population. and environmental domains [49]. Future studies from the
On the other hand, as the other characteristics of the sam- SPEEDY study will address these associations and provide
pled schools are comparable to those of all eligible much-needed information on where and how to target
schools, it appears that we recruited a broadly representa- intervention efforts most effectively [50] in order to
tive sample of schools into our study. improve physical activity and dietary behaviours in chil-
dren.
Of all invited children, 57.0% took part in the data collec-
tion. Response rates may have been affected by the Conclusion
amount of data collected, the speed of the data collection The aim of the current paper was to provide descriptive
(leaving ample time for additional reminder efforts), the data on physical activity and dietary behaviour of the chil-
recruitment method (only contact with children through dren participating in the SPEEDY study. The results show
schools) and previous negative research experiences in that physical activity levels are relatively high, especially
schools. No data is available on those children not taking in boys, indicating that a prevention of decline, rather
part, reducing the potential to establish the representitave- than increasing physical activity levels, might be a modifi-
ness of the study sample. It does appear that girls are over- able focus for intervention development in this age group.
represented in the study and, compared to the Norfolk Although the dietary habits of the SPEEDY children seems
population, that a smaller proportion of obese children to be better than reported in previous studies, a significant
have taken part. Data using school-based assessments of proportion of the children reported consuming limited
obesity levels in Year 6 children across Norfolk in the intakes of fruit and vegetables. Therefore, promotion of
same year indicate that 14.1% are classified as overweight daily fruit and vegetable intake in this age group is war-
and 15.8% obese [48]. It is well-established that non- ranted, possibly focussing on children from lower socioe-
responders to health surveys are more likely to have an conomic backgrounds.
unhealthy lifestyle, and this may have resulted in more
positive prevalence estimates. Our final sample is mainly Competing interests
white with only 3.8% of children from other ethnic back- The authors declare that they have no competing interests.
grounds. Although this is representative of the Norfolk
population, it also reduces the generalizability of our Authors' contributions
results to the wider British population. EvS, PS, AJ, AC and SG were responsible for conception of
the study, study design, set up and data collection and
Strengths and limitations of the SPEEDY study processing. KM coordinated the data collection. AMC, FH
Strengths of the SPEEDY study include the population- and JP participated in the data collection and processing.
based sampling strategy, the large sample size, the use of IH, UE and NW were involved with conception of the
objective and valid measures of physical activity and study and study design. EvS analyzed the data and drafted
anthromopometry and of a frequently used and validated the manuscript. All authors were involved with data inter-
food frequency questionnaire. An intensive measurement pretation, critical revisions of the paper and provided
period meant that measurements were only conducted approval for its publication.
during the Summer term. The advantage of this approach
is the reduced influence of weather on the physical activity Acknowledgements
measurements and its potential confounding effect on the We would like to thank the schools, the children and their parents for their
association with potential correlates. However, it may participation in the SPEEDY study. We also thank everyone who helped
have affected the absolute levels of physical activity and with the data collection and Norfolk County Council Children's Services
for their invaluable input and support. We also thank the HBSC Interna-
diet intake presented. Other limitations include the use of
tional Coordinating Centre for allowing us to use the HBSC dietary ques-
a self-report measure of dietary intake, a suspected higher
tionnaire.
non-response of obese children and the fact that the
demographic profile of the county of Norfolk is not repre- EvS, UE, NW and SG are funded by the Medical Research Council (MRC).
sentative for the whole of Britain. PS, APJ, IH and AC are funded by the Higher Education Funding Council for
England (HEFCE). AMC is funded by a University of Cambridge studentship.
The second aim of the SPEEDY study is to examine indi- FH is funded by an ESRC/MRC interdisciplinary research studentship. The
vidual and collective factors associated with physical SPEEDY study is funded by the National Prevention Research Initiative
http://www.npri.org.uk, consisting of the following Funding Partners: British
activity levels and dietary behaviour in a large population-
Heart Foundation; Cancer Research UK; Department of Health; Diabetes
based sample of British 9–10 year old children. The study
UK; Economic and Social Research Council; Medical Research Council;
is the first European study in children setting out to collect
Research and Development Office for the Northern Ireland Health and
objectively measured physical activity and detailed dietary
Social Services; Chief Scientist Office, Scottish Executive Health Depart-
information together with a wide range of potential corre- ment; Welsh Assembly Government and World Cancer Research Fund.
Page 10 of 12
(page number not for citation purposes)